A learner finishes a Science answer and asks:
“Is this correct?”
That question is understandable.
But there is a more powerful question available first:
“What evidence do I have that this answer meets the success criteria?”
Self-assessment begins when the learner can inspect the quality of their own scientific work before the teacher supplies the verdict.
This guide belongs to the Primary 4 Science Learning Hub. Its job is distinct from generic metacognition and revision planning: this page owns judging a specific piece of scientific work against explicit criteria.
Why Self-Assessment and Rubrics Matter
Modern formative-assessment practice emphasises clear expectations, feedback and learner involvement. In science education, quality matrices and rubrics can help pupils recognise what a strong representation, explanation or investigation looks like, and can support teacher, self and peer assessment.
Useful external reference: Primary Connections | Providing formative feedback.
The current MOE Primary Science syllabus also frames assessment as information that should inform future learning, not merely produce a mark.
Quick Answer: The Self-Assessment Loop
KNOW THE CRITERIA → COMPLETE THE TASK → CHECK EACH CRITERION → MARK THE FIRST GAP → REVISE ONE THING → RECHECK → SUBMIT / RETEST
This is an eduKate teaching routine, not an official MOE marking scheme.
Wait, What? A Rubric Is Not a Marking Scheme
A marking scheme tells an assessor how marks are awarded.
A learner-facing rubric or success-criteria checklist can instead answer:
- What quality should my answer show?
- What scientific relationship must be visible?
- What common weakness should I look for?
- What can I fix before submitting?
The job is quality control, not mark prediction.
Success Criteria Should Be Observable
Weak criterion:
“Good answer.”
Stronger:
- states the relevant scientific idea;
- connects it to the evidence;
- uses the correct direction of cause/effect;
- answers every part of the command;
- uses units where needed;
- does not overclaim beyond the evidence.
The learner can actually inspect these.
Rubric Level 1 | Present or Missing
For younger learners, a binary checklist is often enough.
| Criterion | Present? |
|---|---|
| Answered the command | □ |
| Used correct scientific idea | □ |
| Connected evidence to conclusion | □ |
| Included unit if measured | □ |
Rubric Level 2 | Emerging / Secure
Another simple scale:
- Not Yet — criterion is missing or incorrect.
- Developing — present but incomplete or prompt-dependent.
- Secure — accurate, relevant and independently expressed.
These are teaching labels, not official grades.
Rubric Level 3 | Quality Matrix
A quality matrix describes what increasing quality looks like for one scientific product.
Example: labelled diagram.
| Feature | Developing | Secure |
|---|---|---|
| Labels | some parts labelled | all relevant parts labelled clearly |
| Arrows | present but meaning unclear | direction and relationship clear |
| Scale | looks literal | states if not to scale |
| Scientific purpose | decorative | relationship is visible |
Success Criteria for an Explanation
Question:
“Explain why Plant Q wilted more after many roots were damaged.”
Criteria:
- mentions root function;
- states reduced ability to absorb water;
- connects reduced uptake to wilting;
- does not claim roots absorb ready-made food;
- uses the actual condition in the question.
Self-Assessment Example | Weak Explanation
Answer:
“Plant Q wilted because its roots were damaged.”
Checklist:
- Condition present? Yes.
- Root function present? No.
- Mechanism present? No.
- Outcome present? Yes.
Next repair:
add water-absorption mechanism.
Success Criteria for Heat Explanation
Strong criteria:
- identifies hotter object;
- identifies cooler object;
- states heat transfer direction;
- separates heat from temperature;
- links transfer to temperature change.
This catches wording such as “temperature flowed”.
Success Criteria for Matter Explanation
Question:
“Why does lower water level not mean lower volume?”
Criteria:
- mentions container shape/width;
- distinguishes height from volume;
- states volume remains constant if none is lost;
- uses measured value if supplied.
Success Criteria for Light Diagram
Criteria:
- source identified;
- object/blocker identified;
- screen or eye identified;
- reference distance labelled;
- arrows/lines have clear meaning;
- does not imply rays are physical strings.
Success Criteria for Digestive Route
- mouth;
- gullet;
- stomach;
- small intestine;
- large intestine;
- correct sequence;
- correct function clues.
Success Criteria for a Fair Test
- one main changed condition;
- measured outcome named;
- important comparison conditions kept similar;
- apparatus suitable;
- method repeatable;
- units stated;
- safety considered.
Success Criteria for a Practical Task
- reads task before manipulating;
- sets apparatus correctly;
- measures intended property;
- uses consistent method;
- records immediately;
- includes units;
- responds appropriately to anomaly;
- can explain why controls matter.
Success Criteria for a Table
- clear headings;
- units in headings;
- conditions separated correctly;
- values placed in correct row/column;
- derived values calculated accurately;
- no decorative clutter.
Success Criteria for a Graph
- axes match variables;
- units shown;
- scale sensible;
- points plotted accurately;
- trend described without overclaiming;
- anomalies not hidden.
Success Criteria for a Physical Model
- purpose stated;
- parts have defined meaning;
- scientific relationship is preserved;
- model can be manipulated or explained;
- one limitation is stated;
- decoration does not replace meaning.
Success Criteria for a Project
- focused question;
- safe method;
- relevant evidence;
- milestones;
- bounded conclusion;
- revision after critique;
- reflection on what changed.
Success Criteria for Scientific Discussion
- claim stated;
- evidence identified;
- reasoning explained;
- critique targets the idea;
- revision occurs if evidence demands it.
Success Criteria Must Match the Task
Do not use the same rubric for everything.
A graph rubric should not judge whether a digestive route includes the gullet unless that graph is somehow about digestion.
Criteria should reflect the product being created.
Too Many Criteria Can Defeat Self-Assessment
A 25-item checklist may become another reading burden.
For one Primary 4 task, use the few criteria that most strongly determine quality.
Often 3–6 is enough.
Self-Assessment Before the Model Answer
Important sequence:
- attempt independently;
- check against criteria;
- revise;
- only then compare with teacher/model answer.
If the model answer comes first, the exercise can become copying rather than self-assessment.
The Two-Pass Check
Pass 1 — Science:
- correct model?
- correct evidence?
- correct causal direction?
Pass 2 — Communication:
- answered command?
- units?
- exact nouns?
- bounded conclusion?
Separate passes reduce cognitive overload.
Red–Amber–Green as a Fast Tool
A simple self-rating:
- Green: can meet criterion independently.
- Amber: can do with one cue or familiar example.
- Red: model or skill is missing.
This is a diagnostic teaching tool, not an official grade.
Confidence and Criteria Are Different
A learner can feel confident and still miss a criterion.
A learner can feel uncertain and still produce a complete answer.
Self-assessment should inspect work, not mood alone.
Self-Assessment Example | MCQ
After choosing C, ask:
- Did I identify the target before reading options?
- Can I explain why C fits?
- Can I reject at least one distractor scientifically?
- Did I check unit/reference point?
This goes beyond “I got it right”.
Self-Assessment Example | Open-Ended Answer
Criteria:
- answered every part;
- included mechanism;
- used given evidence;
- stopped when complete.
The learner can revise before seeing the official answer.
Self-Assessment Example | Practical Measurement
After measuring:
- Was zero aligned?
- Was instrument appropriate?
- Was unit written?
- Was method repeated consistently?
- Does the value pass a sanity check?
Self-Assessment Example | Field Observation
- Did I record time and location?
- Did I separate observation from inference?
- Did I identify uncontrolled factors?
- Did I generate one follow-up question?
Self-Assessment Example | Project
- Is the central question still visible?
- Does the evidence answer it?
- Did the pilot reveal a needed revision?
- Does the final product show limits?
Peer Assessment Uses the Same Criteria
A peer should not say:
“I like it.”
Instead:
“Your diagram has all the labels, but the arrows do not show whether they mean movement or cause.”
Criteria make peer feedback specific.
Peer Assessment Is Not Marking Authority
The peer’s job is to compare work with criteria and ask useful questions.
The teacher remains responsible for formal assessment decisions.
Co-Creating Criteria
Sometimes criteria can be built with pupils.
Show two examples:
- one weak;
- one strong.
Ask:
“What makes the second one easier to trust?”
Pupils may identify:
- units;
- clear labels;
- evidence;
- causal link;
- controlled conditions.
Now the rubric is understood rather than merely handed down.
Rubrics Should Describe Quality, Not Personality
Weak criterion:
“Careful student.”
Better:
“Uses the same measurement position for all trials.”
Judge observable scientific behaviour.
Rubrics Should Not Reward Length Automatically
A concise complete explanation can be stronger than a long answer full of unrelated facts.
Criteria should reward scientific completeness and relevance.
Rubrics Should Not Reward Decoration Automatically
A colourful model can still misrepresent the Science.
Criteria should prioritise meaning.
Rubrics and Model Limits
For diagrams or models, include:
“States what the model does not show.”
This prevents over-literal understanding.
Rubrics and Transfer
A piece of work can meet criteria in one familiar example and still fail transfer.
Add a criterion such as:
“Can apply the same model to a changed surface example.”
Rubrics and Independence
Record support level:
- teacher modelled;
- prompted;
- one cue;
- independent.
Correctness alone does not reveal independence.
Rubrics and Self-Correction
A powerful criterion:
“Can identify one weak link and repair it without being given the correct sentence.”
This is a higher-quality learning signal than passive correction.
Rubrics and Feedback
Rubrics make feedback actionable.
Instead of:
“Your answer needs improvement.”
Say:
“Criterion 3 is missing: you stated the result but not the mechanism.”
Now the learner knows what to repair.
Rubrics and Revision Planning
Self-assessment identifies which criterion failed.
The separate revision-planning system then decides how much study time that failure deserves.
Do not merge the two jobs.
Original Self-Assessment Clinic 1 | Heat
Answer:
“The spoon got hot because metal is a conductor.”
Criteria:
- hotter/cooler identified? No.
- heat-transfer direction? No.
- conduction idea? Yes.
- temperature outcome? Partial.
Repair:
add source of heat and transfer direction.
Original Clinic 2 | Matter
Answer:
“The volume decreased because water level is lower.”
Criteria:
- distinguishes height/volume? No.
- uses conservation condition? No.
- states measured property? Partial.
Repair:
wide container changes height but not volume if none lost.
Original Clinic 3 | Light
Diagram has source, object, screen but no distance labels.
Criterion:
reference points must be explicit.
Repair:
label object–source or object–screen distance as required.
Original Clinic 4 | Plants
Answer:
“Roots help the plant.”
Criteria:
- specific function? No.
- question-linked consequence? No.
Repair depends on whether task is anchoring or water absorption.
Original Clinic 5 | Practical
Reading written as:
“58”
Criteria:
- value? Yes.
- unit? No.
- condition? Missing.
Repair:
58°C at 15 minutes.
Success-Criteria Cards
A useful classroom method is to give small cards:
- MODEL
- EVIDENCE
- LINK
- UNIT
- BOUNDARY
Before submitting, the learner places only the cards that are visibly present in the answer.
Missing cards reveal missing work.
One-Criterion Repair
Do not ask the learner to rewrite everything when only one criterion is missing.
Example:
accurate explanation, missing evidence.
Repair only evidence connection.
This preserves what already works.
Stop Rule
Self-assessment should stop when:
- the relevant criteria have been checked;
- one or two meaningful repairs are made;
- the learner can explain why the work now meets the standard.
Endless checking can create anxiety without improving quality.
Common Self-Assessment Errors
- checks against feelings rather than criteria;
- uses too many criteria;
- looks at answer first;
- marks everything green automatically;
- uses rubric as a score prediction tool;
- rewards length or decoration;
- cannot explain why criterion matters;
- rewrites entire answer instead of first gap.
Original Practice Set
Question 1
What is the difference between a learner rubric and a marking scheme?
Question 2
Why should criteria be observable?
Question 3
Why should self-assessment happen before seeing the model answer?
Question 4
What is the benefit of a two-pass check?
Question 5
Why can confidence differ from quality?
Question 6
How can peer assessment become more specific?
Question 7
Why should support level sometimes be recorded?
Question 8
What is a one-criterion repair?
Practice Answers
1. A learner rubric describes quality to guide checking and improvement; a marking scheme allocates marks for formal assessment.
2. The learner needs to be able to inspect whether the criterion is present or missing.
3. Otherwise checking can become copying rather than independent evaluation.
4. It separates scientific correctness from communication details, reducing overload.
5. Feelings do not guarantee that criteria are met.
6. Use shared success criteria to identify the exact missing or weak feature.
7. A correct answer with heavy prompting is different from independent mastery.
8. Repair only the first missing criterion instead of rewriting work that already meets other criteria.
The Self-Assessment Diagnostic
| If the learner… | Likely weak link | Repair |
|---|---|---|
| says “looks okay” | criteria visibility | use 3–5 explicit checks |
| copies model answer | assessment timing | check before reveal |
| checks everything at once | cognitive load | two-pass system |
| rewrites good work | repair precision | fix first failed criterion |
| cannot transfer criteria | surface dependence | apply rubric to new example |
A 30-Minute Self-Assessment Lesson
Minutes 1–5: compare weak and strong examples.
Minutes 6–10: co-create 4 success criteria.
Minutes 11–15: attempt one task independently.
Minutes 16–20: self-check before model answer.
Minutes 21–25: repair one criterion.
Minutes 26–30: transfer criteria to a new task.
What Parents and Tutors Can Ask
- “What are the success criteria before you begin?”
- “Which criterion is strongest?”
- “Which one is missing?”
- “Can you fix only that part?”
- “Do you need the answer yet?”
- “Can the same criteria check a new example?”
Continue Batch 19
- Practical Tests and Performance Tasks Under Real Conditions
- Acting on Feedback, Feedforward and the Next Repair
- Collaborative Investigations, Group Roles and Shared Evidence
The Quiet Return
Self-assessment is not the learner becoming their own examiner.
Make quality visible. Attempt independently. Compare the work with a small number of meaningful criteria. Repair the first real gap. Recheck. Then let the model answer confirm or challenge the reasoning rather than replace it.
